of small Variations of Temperature on Steel Magnets. 339 













Fall in 



Bar. 



Specific gravity at 

 15° -5 C. 



Specific gravity 

 at37°*7. 



Decrease 



in specific 



gravity 



for 

 22°-2 C. 



Specific 

 gravity 

 at 20°, 

 calcula- 

 ted, = G. 



specific 



gravity 

 per 1°C. 



as a de- 

 cimal of 



specific 

 gravity at 

 20°, =e. 



Observed. 



Cor- 

 rected. 



Observed. 



Cor- 

 rected. 



B ... 



7-845 



7-798 



7-882 



7-750 



•048 



7-788 



•000278 



C ... 



7-880 



7-833 



7-896 



7-764 



•069 



7-819 



•000397 



D ... 



7657 



7611 



7-685 



7-556 



•055 



7-600 



•000326 



E ... 



7-806 



7-760 



7-819 



7-688 



072 



7-745 



•000419 



F ... 



7-883 



7-836 



7-899 



7767 



•069 



7-822 



•000397 



Bar. 



G 3 . 



Decrease in 

 magnetic 

 power for 

 1°C. as a 

 decimal of 

 magnetic 

 power at 

 20°, =F. 



i^mount 

 of impu- 

 rity in 

 100,000 

 parts, 

 = P. 



Log P. 



p7. 



FG 3 



e 



FG 3 



B ... 



47148 



000738466 



1320 



3-12057 



265-8 



1252-4 



4-712 



C ... 



47805 



•001338745 



1823 



3-26012 



341-3 



16120 



4-723 



D ... 



438-98 



•000928042 



1325 



312221 



266-6 



1249-6 



4-687 



E ... 



464-53 



•00155040 



1970 



3-29446 



362-9 



1718-8 



4-736 



F ... 



478-64 



•001389101 



877 



294300 



2121 



1674-7 



7-7 





Mean of 

 Extreme 



BCDE=4-714. 



i error (D)=0*57 pe 



r cent. 







The difference in bar F is immense, but is, we think, accounted 

 for by the fact that the bar is not homogeneous. Magnets cut 

 from different pieces of it gave values of (F) varying from '0013 

 to -0020, while the specific gravity varied from 7*771 to 7*836. 

 Without this, the want of agreement in the analysis is sufficient 

 of itself to show that the bar is not homogeneous. 



We think it most likely that the impurity, when heaped up at 

 the poles or centre of the magnet, would have a very different 

 effect from what it would have when evenly distributed. 



To sum up, for any homogeneous steel bar hardened at dull 

 redness, the law which seems to hold with some exactitude 

 is : — 



Z 2 



